Spindle statistics as biomarkers of physiological transitions

Determine whether deviations in the statistical distributions of spindle properties—specifically spindle durations, peak amplitudes, and inter-spindle intervals extracted via Empirical Mode Decomposition of EEG signals or generated by the two-dimensional Ornstein–Uhlenbeck process—can serve as biomarkers indicating physiological transitions such as sleep stage onset, emergence from general anesthesia, or pathological states.

Background

The paper models transient spindle-like oscillatory bursts in EEG as trajectories of a two-dimensional Ornstein–Uhlenbeck process and introduces an Empirical Mode Decomposition-based segmentation algorithm to quantify spindle features. The authors analyze distributions of spindle durations, amplitudes, and inter-spindle intervals and relate them to process parameters.

In the discussion, the authors propose using these statistical descriptors to potentially identify physiological transitions (e.g., sleep stages or anesthesia changes), but explicitly note that this remains an open question.